January 15, 2014

Pew Research increases share of interviews conducted by cellphone

Pew Research Center is making an important change in the way that we survey Americans by telephone. In the coming months, 60% of interviews in our national polls will be conducted via cellphones and 40% on landline phones. Over the past year, the ratio has been half cellphone, half landline.

We’re doing this because more Americans today no longer own landlines and rely only on cellphones. When we conduct our random-digit-dial phone surveys, we want to ensure that we’re reaching a sample of the population that is representative of the public. This shift has been gradual. When Pew Research started calling cellphones in 2008, they comprised 25% of our respondents. We increased that percentage to 33% in December 2009, to 40% in July 2011 and finally to 50% in January 2013.

According to the most recent report from the CDC’s National Center for Health Statistics, four-in-ten U.S. adults owned only a cellphone in 2013. Hispanics, African Americans, younger adults and the poor are more likely to use only a cell phone. By increasing the number of cellphone-only respondents, we can also ensure that we’re more accurately representing these groups.

In addition to capturing more of these “cell onlys”, this change also yields a better representation of the nearly one-in-five people (about 18%) who have both kinds of phones but rely primarily on their cellphones. These “dual users” who are contacted for surveys via cellphones are also demographically distinct and need to be properly represented in national samples. Among the people we interview on cellphones, about half only use cellphones and the remainder are “dual users.”

Cellphone interviewing costs more and takes more interviewer time than those conducted on landlines, but the improvement to the quality of our surveys is clear. For example, even with half of our interviews being conducted on cellphones, just 19% of respondents in our current surveys are under age 35; nationally this group comprises 31% of the adult population. We project that the share of young adults represented in our surveys will grow to 22% with our shift to more cellphone interviews.

Reaching the young adult population and ensuring they are properly represented in national surveys has always been a challenge for survey researchers. Our adjustments will still take us far short of the target of 31%, which means we will continue to make statistical adjustments to bring our samples into alignment with the population – not only on age, but other demographic characteristics through a process known as “weighting.”

And moving forward, we will continue to make adjustments to our survey methodology to address the changing technological landscape. Some have already suggested that in today’s mobile society, landline interviewing could be dropped entirely since nearly every American has a cell phone. But the current research suggests that we are not there–not yet, at least.

Some 7% of Americans have landline phones with no wireless access in their homes, and a significant number of people who have cellphones in their home are still not viably reachable over that device, either because another individual is the primary user, or because their cellphone is off most of the time, except for in emergencies. For the near future, we expect that a blend of cellphone and landline interviewing will continue to provide the broadest cross-section of the American population.

More information about our polling:
Why am I never called to be polled?
What good are polls?
Do pollsters have a code of ethics?
Why don’t you just conduct surveys on your website?
Other frequently asked questions

Topics: Research Methods

  1. Photo of Kyley McGeeney

    is a senior research methodologist at Pew Research Center.

  2. Photo of Scott Keeter

    is a senior survey advisor at Pew Research Center.


  1. Julie Pokela3 years ago

    How do you obtain your cellphone sample list? As there is no complete list of cell phone numbers or random digit dialing option for cell phones, aren’t you using a convenience sample?

    1. Scott Keeter3 years ago

      You are correct that we don’t have a list of cellphone numbers — no one does. But we do have a list of all telephone exchanges and blocks of numbers assigned to cell phones, and that allows us to draw a sample that is representative of the full population of cell phones. We generate the actual telephone numbers we call by adding random numbers to the end of the sampled blocks of numbers, just as we do with landline numbers. That process gives every cell phone in the country, including newly assigned numbers and prepaid phones, an equal chance of being included in our sample. That’s why we can be confident that our samples are representative of all cell phone owners.

  2. Russell Davis4 years ago

    You do a terrific job.
    I’m writing a paper on inequality, poverty, and hunger in America in 2014 and have used your information extensively.
    Thank you.
    Russ Davis